Detection of plant detriments

Abstract

A computer-implemented method and system for detecting plant detriments includes receiving images of a plant area captured by at least one camera. For each received image, the method includes dividing the image into multiple image zones and processing the image by applying a sequence of machine learning plant detriment processing models to each zone. Plant detriment information in the plant area is determined based on the processing. The processing may include applying classifiers to detect plant detriment artifacts in image zones, utilizing information from previously processed zones. An alert may be issued or a report provided based on the determined plant detriment information. The system includes cameras for capturing plant area images and a processor configured to divide images into zones, apply machine learning models to the zones, and determine plant detriment information. The plant detriments may include pests or diseases.

Dependencies are inferred from the claim text.